A Quick Emergency Response Model for Micro-blog Public Opinion Crisis Based on Text Sentiment Intensity
Mingjun Xin, Hanxiang Wu, Zhihua Niu · Journal of Software · 2012
Abstract—On the basis of discussing the information spreading mechanism under Internet environment, we have studied on how to build a public opinion monitoring model according to the semantic content or text mining in recent years. A micro-blog public opinion corpus named MPO Corpus on the content of micro-blog information as a test data set has been constructed by our research team. In this paper, it proposes a quick emergency response model (QERM) for micro-blog public opinion crisis oriented to Mobile Internet services. Firstly, it describes the micro-blog cases and emergency response plan library using web ontology language (OWL), which makes the transitive logical reason capacity among micro-blog subjects, micro-blog cases and emergency plans. Secondly, it proposes an algorithm to calculate the sentiment intensity of micro-blogs from three levels on words, sentences and documents based on HowNet Knowledge-base respectively. Thirdly, we continue to study on how to update cases under the subjects and quick response processes for micro-blog case base. Finally, we design a test experiment which shows some merits of QERM in time, which basically meets the quick emergency response demand on the micro-blog public opinions crisis under Mobile Internet environment. Thus, it will provide more efficient support to the government and related monitoring departments involved with the public opinions crisis. Index Terms—public opinion crisis; sentiment intensity;